The present invention relates generally to fault detection and fault isolation in an optical network. More specifically, the present invention relates to a method and system for providing fault detection and isolation in an optical network having amplifier nodes.
Optical data networks typically include a plurality of nodes linked by optical fibers into a network. The network may be one of several common topologies, such as a linear chain network, an optical star network, or an optical ring network. Optical networks are also classified by the geographic size of the network, with wide area networks (WANs) and metropolitan area networks (MANs) being of increasing interest for providing high bandwidth network data links to and from corporations and LAN campuses, for example.
A popular optical network topology for MANs is an optical ring. As shown in FIG. 1, an optical ring network 100 typically comprises a sequence of network nodes 105, at least one primary optical fiber path 110, commonly known as “working fiber” coupling data between the nodes. Optical networks transport large flows of information such that system outages of even a few seconds can cause the loss of huge quantities of information. This is especially true for wavelength division multiplexing (WDM) and dense wavelength division multiplexing (DWDM) optical networks, which simultaneously transmit data in a plurality of optical channels, with each channel comprising a different optical wavelength.
The reliability of an optical network is an important design consideration. Optical networks can fail due to several different mechanisms. Line failure is commonly defined as a fault in the ability of light to be transmitted between nodes along a working fiber, i.e., there is no light coupled into the node because of damage to the optical fiber. Additionally, a line failure can occur at or near the interface of the fiber and a node. For example, the optical fiber may not be properly inserted into the node. Additionally, a failure of an optical interface element may be optically equivalent to a line fault if it results in a total loss of signal at all channel wavelengths to all downstream components. For example, a fault in the optical interface element receiving signals from the fiber that results in a complete loss of signal to all subsequent optical elements within the node is equivalent in effect to a fault in the fiber. An electrical equipment failure is commonly defined as a failure in one or more electrical or electro-optic modules in the node. These include optical amplifiers, multiplexors/demultiplexors, transponders, and other elements used to amplify, frequency shift, or to add or drop individual channels or bands. An electrical equipment failure may result in a loss in all channels, but may more commonly result in only a limited number of channels being dropped.
Optical networks typically employ several different approaches to permit network service to be rapidly restored in the event of a fault. Referring again to FIG. 1, optical ring networks typically include at least one protection fiber 115 between each node 105. The protection fiber 115 provides an alternate path for optical data in case the primary optical fiber 110 becomes broken or damaged along a portion of its length. Additionally, the protection fiber facilitates the routing of data to bypass a defective node 105 via a path in the protection fiber. In the case of a unidirectional path-switched ring (UPSR) the working fiber and the protection fiber commonly carry information in opposite directions, e.g., data is commonly transmitted in the working fiber in a clockwise direction and in the protection fiber in the counter-clockwise direction. Bi-directional path switched rings (BPSR) permit traffic along the ring to be carried in both directions via two or more working fibers and two or more protection fibers.
FIG. 2 is an illustrative diagram of a UPSR ring 200 operating with working and protection fiber path links intact. For the purposes of illustration, an optical data path is shown between the tributary interfaces of two nodes, NE1 and NE2. As shown in FIG. 3, in the event of a fiber break the working traffic is switched to the protection fiber in order to maintain the data link between the tributary interfaces of nodes NE1 and NE2. This is performed using optical line switching elements (not shown in FIGS. 2 and 3) within a node in order to optically switch the path of the optical signals. Note that a complete failure of one or more electrical elements within node NE1 or NE2 could also break the flow of data. Consequently, nodes NE1 and NE2 typically include redundant electrical and electro-optic elements that can be switched into use in the event that one or more electrical elements in the node fails. This is commonly known as equipment switching.
As shown in FIG. 1, a network management system (NMS) 120 is typically used to regulate the action of the nodes 105 in the event of a line failure or an equipment failure in order to restore network service. The NMS 120 typically comprises a central workstation computer receiving electrical signals corresponding to the optical strength of every optical channel transmitted through each active line of each node. The NMS 120 is typically programmed with a list of rules or procedures for handling different types of failures. Multi-channel optical-to-electrical-to-optical (OEO) detectors (not shown in FIGS. 1–3) in each node can be used to measure the signal strength of each channel entering or leaving the node. This permits the NMS 120 to determine if a channel has been dropped. If the NMS 120 determines that a channel has been dropped in a particular node, the NMS may instruct the node to perform an equipment switch of a component in the path of the dropped channel likely to have failed. The NMS 120 monitors the activity of all of the nodes, determines if a change in traffic occurs, makes a decision whether a line fault or equipment fault has occurred, isolates the fault to a particular node or fiber path, and issues appropriate commands to all of the nodes to perform one or more equipment switches or line switches to restore network traffic.
While the network management system shown in FIG. 1 improves the reliability of network 100, the inventors of the present application have recognized that it has several substantial drawbacks, particularly in regards to high performance metropolitan area networks. First, it can take a significant length of time for a central computer of a NMS 120 to determine an appropriate course of action due to the cumulative time delays of the system. There are finite response times for each OEO to measure the signal strengths of each optical channel to determine if a channel is dropped. There is also a significant propagation time for channel status signals to reach the central computer of NMS 120. This propagation time includes the time delay for short-haul Ethernet cables coupled to the node along with the time delays of the long-haul data link (e.g., a telephone line) to the central computer, which may be located several kilometers away from an individual node in network 100. There is also a time period required for the central computer to assess the state of each node and to make a decision. Still yet another time period is associated with the time delay required to transmit control signals from the central computer of NMS 120 back to each node via Ethernet and long-haul connections. There is also a time delay associated with the circuitry at the node that is used to implement a line switch or an equipment switch. In a conventional MAN system 100 the total elapsed time between the detection of a failure and a line switch or equipment switch being implemented can exceed 0.1 seconds. One industry standard that has evolved is that a communication disruption lasting more than 50 milliseconds constitutes a network outage, i.e., tributary networks receiving and transmitting data via network 100 are designed upon the assumption that optical network 100 has outages of less than 50 milliseconds. Network outages in excess of 0.1 seconds may therefore cause an irreparable loss of data to a tributary network.
Another drawback of network 100 is that the NMS 120 can be comparatively expensive to implement. The central computer is often implemented as a high performance work station, which is comparatively expensive. Another substantial cost is associated with the OEO modules used to measure channel strength in each node 105. OEO modules increase with cost as a function of the number of optical channels that they are capable of analyzing. Additionally, the cost of each OEO module tends to increase with the data capacity of each channel since faster optical and electrical components are required for high data rate channels. Advances in DWDM technology now permit thirty or more high data-rate channels to be implemented in an OEO module. This results in a corresponding increase in the cost of the OEO modules compared with first generation WDM designs having three to five moderate data-rate channels.
Another drawback of network 100 is that it may provide insufficient information to isolate electrical equipment failures for later repair. The increase in the number of channels in DWDM systems has led to multistage node designs having several stages. The stages commonly include various combinations of band pass filters, channel filters, wavelength shifters, optical amplifiers, multiplexors, and demultiplexors. Each stage, in turn, may host one channel, several channels, most of the channels, all of the channels, or frequency shifted versions of the channels, depending upon the function of the stage. A single OEO module is typically insufficient to determine the element within a node that has dropped a channel. Consequently, several OEO modules may be required for fault isolation, further increasing the cost of the system.
Furthermore, an optical fiber span between two nodes in network 100 is often too long to allow optical signals to be received within an acceptable level of confidence. Generally, the length of the optical fiber span is directly proportional to signal degradation, i.e., the greater the length of the optical fiber span, the more serious the signal degradation becomes. Consequently, one or more amplifier nodes may be added along an optical fiber span to amplify the optical signals to improve signal reception. The primary function of the amplifier nodes is simply to amplify the optical signals propagating along the optical fiber span. The use of amplifier nodes, thus, allows the length of an optical fiber span to be extended. Extending the length of an optical fiber span, however, causes at least one drawback. Since the length of the optical fiber span is increased, the time it takes for an optical signal to travel from one node to another is also increased. As a result, when a line fault involving an amplifier node occurs, the protection switch time needed to restore line traffic is increased. This increase in protection switch time may prevent a network from achieving the desired sub-50 ms protection for line traffic, thereby causing more network outages to occur.
Therefore, it would be desirable to provide an improved system and method for performing fault detection, isolation, and network restoration in an optical network having amplifier nodes.